Description
- Abstract:
- The High-Luminosity Large Hadron Collider (HL-LHC) will dramatically increase the volume collision data, giving significant challenges to current computing and storage capabilities. In this study, we explore the application of lossy compression techniques to quark and gluon particle-level jets using the Baler algorithm. We investigate three different autoencoder-based models: Conventional Auto-Encoder, Single-feature Auto-Encoder, Forked-network Auto-Encoder, to perform lossy compression on jet constituent information. To further evaluate the preservation of physical information after compression, Particle Flow Networks (PFNs) are employed for jet tagging on the reconstructed jets, and the relationship between the compression ratio and the area under the ROC curve (AUC) is expected to be analyzed systematically. Our results show that although compressed representations maintain good jet tagging performance, challenges such as information loss due to zero-padding and discrepancies in jet mass distributions remain.
- Notes:
- Thesis (Sc. M.)--Brown University, 2025
Access Conditions
- Use and Reproduction
- All rights reserved
- Rights
- In Copyright
- Restrictions on Use
- All Rights Reserved
Citation
Dong, Shutong,
"Studies of the Effects of Lossy Compression on Hadronic Jet Reconstruction and Classification"
(2025).
Physics Theses and Dissertations.
Brown Digital Repository. Brown University Library.
https://doi.org/10.26300/08pj-9466
